Discussion of "A possible generating algorithm of the Voynich manuscript"

[nablator]

Noob question: if your words are all very short, won't this increase the likelihood of lower values?

Looking at possible confounding variables is an essential part of the analysis of correlations.

The variation in average word lengths alone might explain some if not all of the variation of the average edit distance between words of different lines of the same page.

For example TT Herbal A:

   

Horizontal axis: vertical distance in lines (inside pages, not across different pages)
Vertical axis: average absolute value of the difference between lengths of all couples of words from these lines

Again, the values on the far right of the graphs are not significant because of the reduced size of the sample.